| Passive Target Tracking in Non-cooperative Radar System Based on Particle Filtering |
| |
| 作者姓名： | 李硕 陶然 |
| |
| 作者单位： | Department of Electronic Engineering, Beijing Institute of Technology, Beijing 100081, China |
| |
| 基金项目： | 中国科学院资助项目
,
教育部高等学校骨干教师资助计划 |
| |
| 摘 要： | We propose a target tracking method based on particle filtering（PF） to solve the nonlinear non-Gaussian targettracking problem in the bistatic radar systems using external radiation sources. Traditional nonlinear state estimation method is extended Kalman filtering （EKF）, which is to do the first level Taylor series extension. It will cause an inaccuracy or even a scatter estimation result on condition that there is either a highly nonlinear target or a large noise square-error. Besides, Kalman filtering is the optimal resolution under a Gaussian noise assumption, and is not suitable to the nonGaussian condition. PF is a sort of statistic filtering based on Monte Carlo simulation that is using some random samples （particles） to simulate the posterior probability density of system random variables. This method can be used in any nonlinear random system. It can be concluded through simulation that PF can achieve higher accuracy than the traditional EKF.
|
| 关 键 词： | 雷达 滤波 目标跟踪 表面辐射 |
| 文章编号： | 1673-002X（2006）01-0053-04 |
| 收稿时间： | 2005-02-04 |
Passive Target Tracking in Non-cooperative Radar System Based on Particle Filtering |
| |
| LI Shuo,TAO Ran.Passive Target Tracking in Non-cooperative Radar System Based on Particle Filtering[J].Journal of China Ordnance,2006,2(1):53-56. |
| |
| Authors: | LI Shuo TAO Ran |
| |
| Abstract: | We propose a target tracking method based on particle filtering(PF) to solve the nonlinear non-Gaussian target-tracking problem in the bistatic radar systems using external radiation sources. Traditional nonlinear state estimation method is extended Kalman filtering (EKF), which is to do the first level Taylor series extension. It will cause an inaccuracy or even a scatter estimation result on condition that there is either a highly nonlinear target or a large noise square-error. Besides, Kalman filtering is the optimal resolution under a Gaussian noise assumption, and is not suitable to the non-Gaussian condition. PF is a sort of statistic filtering based on Monte Carlo simulation that is using some random samples (particles) to simulate the posterior probability density of system random variables. This method can be used in any nonlinear random system. It can be concluded through simulation that PF can achieve higher accuracy than the traditional EKF. |
| |
| Keywords: | passive radar system target tracking particle filtering Particle Filtering Based Radar System Target Tracking Monte Carlo simulation accuracy used simulate posterior probability density system variables particles sort statistic random Kalman filtering optimal resolution |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
